My Trading System: A Brief Summary

Brett N. Steenbarger, Ph.D.

www.brettsteenbarger.com
Note: This was written December 1, 2003 and reflects the system as of the date of
writing. Modifications to the system and its implementation are ongoing and are
followed in the Trading Psychology Weblog.
This article describes a non-mechanical, rule-based intraday trading system for the ES
(S&P E-mini) futures, though it can easily be adapted to trading the SPY ETF or the full
sized SP futures. There are three major inputs to the system: TREND, INSTITUTIONAL
ACTIVITY, and MOMENTUM. The basic logic of the system is to identify the shortterm trend as bullish, bearish, or neutral and predicate intraday entries on the behavior of
the Institutional Buying/Selling and Momentum during the day. This is explained below.
The specific indicators that assess trend, institutional activity, and momentum are
summarized daily in the Weblog.
There are two trading modes for the system:

Short-term trend following. This is based on research that shows that, over the
short-run, trending price movements of a threshold vigor are likely to continue in
their present direction, while over the longer-run, they are apt to reverse.

Breakout. This is based on research that shows that, when there is no established
trend and the market is trading within a range, a break above or below that range
is likely to continue in its present direction if it reaches a threshold vigor, and is
likely to reverse if it does not reach this critical mass.

In an established bullish or bearish trend, a modeling approach is used to generate an

expectable countertrend movement. When this movement occurs, it forms the basis for an
entry. Thus, the system buys the market on pullbacks during a short-term uptrend and
sells the market on bounces during a short-term downtrend. By looking at markets from
the recent past similar to those of the current market, it is possible to generate a profile of
past bounces and pullbacks to provide a template for when to enter the current market.
The system is non-mechanical because these profiles are always changing, in accordance
with recent market action. An expectable pullback or bounce in a high volatility market is
not the same as in a low volatility market, for example.
When the trend reading is neutral, the system goes into breakout trading mode. To buy or
sell an upside or downside breakout, the market must move to a new high or low, register
breakout readings on the NYSE TICK (to demonstrate a high degree of participation
among stocks), and demonstrate a plurality of individual stocks participating in the move

by registering new short-term highs or lows. I maintain a basket of stocks that I follow
closely intraday. If half or less of the stocks in the basket fail to register a fresh new high
or new low during a candidate breakout move, the system will not take the move. (If the
Trend is going the other way, the system could even fade the move per the rules above).
Only when price, TICK, and new highs/lows are moving in concert does the system take
a breakout move.
Exits are primarily time based with two overrides. The same modeling process mentioned
above is used to generate a profile for the duration and extent for expectable moves,
given similar markets in the past. Once the duration or extent is reached and there is a
very short-term loss of momentum (assuming the trade is profitable), the position is
exited. The first override is if the position is profitable and volume and/or volatility
expands in the direction of the trade by a threshold amount, an immediate profit is taken
as soon as very short-term momentum wanes. The second override is a stop loss, which is
averages 3 ES points, but is adjusted for volatility. This stop loss is moved to breakeven
once the position is profitable by a threshold amount (averaging 1.5 points, but also
adjusted for volatility.
The general goal is to exit under one of three conditions:

When the trade is profitable and traders begin piling into the move, raising
volume and volatility to unsustainable levels.

When the trade is unprofitable from the outset and it is important to limit losses.

When the trade is neither profitable nor unprofitable, but isnt moving in the
expected direction in the expected time frame. The time-based stop allows the
trader to scratch the trade before it hits the stop loss and becomes a loser.

My informal observation is that the third stop condition is the most important. If I am
following my rules and trading in the direction of trend, institutional activity, and
momentum, my trade should go profitable in a relatively short amount of time. Once a
threshold time period is reachedand research is needed to define what that period is
the odds are no longer in the trades favor (partly because the market momentum is being
lost). Learning to exit trades before they are losers is a skill critical to maintaining a
favorable P/L.

Brett N. Steenbarger, Ph.D. is an Associate Professor of Psychiatry and Behavioral

Sciences at SUNY Upstate Medical University in Syracuse, NY and a daily trader of the
stock index futures markets. He is the author of The Psychology of Trading (Wiley, 2003)
and coeditor of the forthcoming The Art and Science of the Brief Psychotherapies
(American Psychiatric Press, 2004). Many of his articles on trading psychology and
daily trading strategies are archived at his website, www.brettsteenbarger.com.